Report of IREP-CESP Joint Workshop on The Transformation of Market Research
Paris, June 27th 2013
The business of market research is experiencing a turning point due to two concomitant changes: a change in the behavior of individuals (behavioral change) and a change in the technical means to analyze this behavior (technological change).
Based primarily on their willingness to participate, individuals are more and more difficult to query these days. The proliferation of surveys by phone, email, Internet, in public spaces and retail outlets, all combined with the end of the model of individual (a housewife, under 50 years, always available and happy to answer) has resulted by a steady decrease in the rate of recruitment. For example in a French study carried out continuously by telephone over five years, the rate of recruitment steadily decreased from 22. 5% in 2009 to 19. 2% in 2010, 18. 3% in 2011, 18. 9% in 2012, and 17. 9% in 2013… A continuous decline to the point of asking the question of disappearance of telephone surveys in a few years! And there is no signal that what is true for the phone today will not be true for email and Internet tomorrow!
In addition to this proliferation, the life of an individual is so fragmented that it is more and more complicated for him or her to remember what media he or she was consuming, on which device, and at which time. As a result, surveys and panels (in particular diary-bases ones) become less and less accurate in fully capturing the life of individuals.
The advent of new technologies, such as Big Data and social networks, have allowed new, small, and more agile players to enter the market with solutions, methods and costs that often prove more relevant. Before the Big Data era, data was a scarce commodity. With Big Data, it has become abundant and it will be even more abundant with the Internet of Things when each object will have its IP address, its sensors, its connection to bigger “data centers” where not only data will be stored but equally powerful software to analyze, visualize, and transform this data. Google, Amazon, Microsoft, and Rackspace already make available to the new entrants the same infrastructure they offer to the traditional big players: Salesforce. com, SAS, SPSS, QlikView, and Tableau Software are now available as a service delivered over the cloud on demand and for small fees.
Analyzing these two changes and their impact on market research was the aim of the workshop that IREP and CESP have jointly organized in Paris on June 27th 2013. Different professionals were invited to present their views: academics, auditors, advertisers, researchers, and high-techs.
This report is a summary of the salient points noted from the speakers’ presentations followed by a synthesis of the lessons about the future of market research… For sure, it will be data-rich high-tech driven.
Zysla Belliat, President of IREP and Charreton Bied-Denis, Managing Director of CESP, have each in turn presented technological innovation as the trigger of the two changes. Technology within the reach of consumers changed their behavior towards advertising, media, loyalty, and more generally towards the products and services they consume. The new technology also changed the way to track, analyze, and anticipate consumer behaviors.
For Zysla Belliat, market research studies must adapt to these two changes. To begin, she recommends starting by understanding that now the questionnaires and samples are now only tools among others. On the one hand, the consumer has become fugitive, recalcitrant towards protocols, and present on several “devices”. On the other hand, the new technologies, especially Big Data, provide new tools for the collection, processing, analysis, and delivery of data. The two changes combined have a big impact on the way studies are conceived, planned, and executed, all in an economic context than forces the community to go faster and cheaper. Zysla Belliat calls research professionals to innovate and evolve, but without sacrificing the rigor and quality that make their mark, their raison d’être.
For Denis Bied-Charreton, being perpetually connected, individuals have become producers of data through their actions on all the “devices” available to them: TVs, game boxes, PCs, smartphones, and tablets.
Passive data is now available in mass but can it be a substitute for all or part of the data collected in a declarative and conscious manner? He asked. Should we abandon samples in favor of mass data? He added. Is this the end of the questionnaires and the beginning of sensors? He added again. For Denis Bied-Charreton, studies need to adapt their analysis or build new ones but they first must go back to data, its value, and its quality as priority #1, which he summarizes with the phrase “back to basic, regardless of the data source. “
Presentations made by Zysla Belliat and Denis Bied-Charreton show that the “technologication” of market research and the availability of mass data bring challenges not only to professionals but also to academics: new protocols to collect, process, and analyze data from varied sources… A real research program for academics in behavioral economics and data science.
Anne-Marie Dussaix, Honorary Professor at ESSEC, was more interested by the expected quality from this technology-led turning. To Anne-Marie Dussaix, relevance, accuracy or completeness, timeliness, accessibility, interpretability and coherence and comparability of data collected in mass are now more than ever essential to ensure the quality of a survey.
The growing amount of data should not hide the fact that 21% of the French population does not have an Internet access, more and more households have completely abandoned their fixed lines, and that a growing number of households do not appear on the public listings (they frequently change operator). . . To Anne-Marie Dussaix, a good study should not put priority #1 to quantity, but to quality instead.
She sees the risk of quality of a study as the consequence of shorter deadlines, smaller budgets, lower response rates, and changing behaviors. To reduce this risk, Anne-Marie Dussaix recommends the use of injection and fusion, two techniques that have long been used in media audience studies. An injection method replaces the missing responses in a questionnaire with responses from similar respondents who answered the questionnaire. A fusion method, as its name suggests, combines the results of two or more complementary studies.
Pascale Carle, Director of Research and Prospecting at Auchan, presented the expectations of retailers. Big Data specialists promise retailers a new and rich era; a wealth the new entrants compute from data consumers generate through merchant cards, outlets, websites, travel agencies, and insurances. Meanwhile, traditional suppliers continue their “business as usual”: sampling, questionnaire, collection, processing… For Pascale Carle, there is a large gap between two value propositions. While Big Data opens new horizons and new opportunities, they often come with answers that are not actionable. Big Data tools do not yet help much the retailers. She gives some illustrative examples of this gap: What does the word “bravo!” mean on a social network? What diagnosis to make if this “bravo!” is taken, multiplied again and again on the Web? Is this a sign of total satisfaction or is it rather a sign of total dissatisfaction? What does the word “Auchan” mean? For Pascale Carle, what Auchan customers or fans have written has a greater value than the number of times they have written the word “Auchan”.
She invites market research professionals to quickly integrate Big Data tools in their data collection and analytics platform and add them the know-how they lack: asking the right questions, looking for the right information. What is of the highest value to Pascale Carle are tools to help her forecast sales and/or mix shelves based on weather conditions, according to the day, prices, sporting and cultural events, history of buyer navigation, tweets, likes, and payments…Pascale Carle’s call for action sounds like an invitation for traditional players to join their efforts with the new entrants to design, develop and propose solutions, combining the expertise of the firsts with the technology of the seconds… A call for a data-rich, technology-based research era!
Jean Thibaud, Director of Research and Segmentation at SFR, presented another fresh view coming from actors outside market research: the big consulting firms such as McKinsey, Accenture, Deloitte, BCG and Bain & Company. These firms bring a new vision…The business that was data-driven is now action-driven. Data is everywhere (in logs, tweets, emails, recommendations…) but its true value is in the contextualization, comparison, and merge, and transformation into an actionable plans. Accenture defines this migration of as the Age of Aggregation.
As Anne-Marie Dussaix of ESSEC and Pascale Carle Auchan before, Jean Thibaud asks to focus on quality, on relevance, and not only on volume. A large amount of data does not mean that the information sought is inside. For example, 15 to 20% of the French population is off the radar of Big Data tools. They do not use Facebook, Twitter, and alike but they also have a voice, an opinion, something to say about their consumption. A real challenge for researchers is to get the voice of these silent individuals.
Yannick Carriou, Global CEO at Ipsos MediaCT, presented a general research firm’s point of view. For Yannick Carriou, media studies are out of control due to the flood of new technologies within the reach of individuals. And collecting audience measurements based on memorization of individuals is no longer relevant. Media consumption is split between several “devices” (radio, TV, PC, smartphone and tablet). Each of the devices supports different medias at once. Reading newspapers on smartphones and tablets, watching movies on PCs and tablets. But the real paradigm shift is perhaps one of the new ways of purchasing advertising space in the digital universe. Programmatic buying, based on the analysis of the digital track of individuals, the implementation of complex algorithms, and stirring of millions of data, coupled with auction systems will affect the television one day, as piloted by SKY in the UK, to send campaigns based on known characteristics of the home directly.
He cites new entrants from the high-tech sector (Rocket Fuel in the U. S. , Criteo in France) that bring new ways, new metrics, new methods completely unknown to market research companies, to media agencies, and to their clients. The intelligence of data will grow strongly, and questionnaires will only be a part of the global picture. And when they are used, they will go shorter. Google even offers questionnaires with two questions!But they combine answers of these two questions with thousands or even millions of data collected by other means such as Google search engine, Gmail mail service and Google+ social network. The era is the passive measurement (via smartphones, tablets, and in the near future Google Glass) and artificial intelligence to collect, store, fuse, and comprehend large volumes of data.
In this fragmented world, the implicit or explicit fusion of information coming from various sources will be the rule. As far as the quality data is concerned, CESP and other quality monitoring and control bodies can analyze the quality of raw data and compare it with a quality reference, but the quality of aggregated data and “statistical glue” between heterogeneous sources is still a large open question. And here, the new entrants must be part of the solution.
Benoît Cassaigne, Executive Director at Médiamétrie, presented the perspective of an audience measurement actor (radio, TV, and Internet). Benoît Cassaigne summarizes the new context of media by a formula: digital world = bloated world… We are in a bloated world of offerings, equipment, screens, and data connections. In a few years, the number of TNT channels in France increased from 7 to 24. The number of Internet access has doubled. The number of set-top boxes is 4 times bigger. The new media environment brings with it new challenges for professionals in the measurement.
First challenge: Scope. Now any new media study must blend Panel data (individuals) with Big Data (devices). It must also consider multiple screens… TV is no longer only watched on TV… Radio is no longer only listen to through radio. And processing is no longer executed periodically but increasingly in real time. Listening, viewing, and browsing data should be available all the time and for all the devices.
Second challenge: Quality. Unlike Panel data, we have no standards for Big Data. Logs, tweets and alike are not all certified, and do not always represent the total universe studied… Big Data doesn’t equal Big Picture!
Third challenge: Myopia, which Benoit Cassaigne defines through a telling example. Extrapolating video time spent on YouTube one hand and on TF1 on other hand, we can assume that the total of 96,527,000 hours of viewing per month for YouTube is superior to 50 minutes per day for TF1! The extrapolation is obviously nonsense!
As for the future of media research, Benoît Cassaigne considers the use of multi-channels to collect data (mobile phones, the web), the fusion of data coming from Panels with Big Data, all with more constrained budget and in a shorter time. . . In one sentence “do more with less. “
Thibaut Munier, CEO of 1000mercis, brought the viewpoint of a high-tech company. Thibaut Munier defines Big Data as the enriching of the customer database at each contact of the latter with the brand, company, product, or service: visit to a web site, access to a social network, click on an ad banner, response to an email…
To Thibaut Munier, Big Data expands the scope of market research. Not only studies can be customer-centric, they can also be site-centric or ad-centric. With Big Data, segmentation is not based on socio-professional criteria or previous purchases; it is based on the behavior of the individual: search, compare, purchase/abandon, and advice/recommend… He cites the example of real time bidding, a new sort of algorithms, which can track Web users, monitor the sites they visit and detect which products they were interested in. If the user moves from one site to another, the algorithm will feature ads related to items he had viewed on the first site. The purchase of advertising space on the Web has become real time, depending on the site, the day, major events in progress, the visitor, and so on. Media planning has become to advertising what the high frequency trading is to finance. Already appeared marketplaces called “Ad Exchanges” to refer to their equivalent in Finance, the “Trade Exchanges. “Without knowing it, without always being aware, web pages change for you and you change web pages according to what you do… Welcome to the Web that learns and adapts to all and in real-time!
The most important thing is not whether we live in an evolution or a revolution. The most important is to remember that so far Information Technology (IT) was considered as a main cost center or a support function. Now IT has become increasingly a differentiator and key driver for the actors who master it at best. Digital technologies have become the DNA of a new generation of market research where the consumer is changing, global, recalcitrant, both consumer and producer of data, battling to keep his or her private life invisible to others, in particular to market researchers and advertisers.
We focus a lot on data but with Big Data and social networks, data is no longer a rare commodity. As noted by Jean Thibaud and Yannick Carriou, traditional market research is challenged, new players are already present (high-tech companies on the one hand, consulting firms on the other), new skills are required…It is not unlikely to see a few years from now completely unknown companies in the “Honomichl Global Top 25, 50 or 100.” The next leaders will be those that incorporate the expertise of market research, the computational power of Big Data and the art of presenting insights and actionable recommendations.